A year after Spotify first announced its plans for dealing with a mass influx of AI-generated beats into streaming apps, it says it’s doubling down on “putting human artistry first.”
Related video above: The call to rein in AI is coming from inside the industry
“This year alone, we’ve taken more than 50 million slop and spam tracks out of recommendations,” a press release said. “This is in addition to more than 75 million spammy tracks we’ve removed from our platform altogether for serious or repeated violations of our rules.”
The report doesn’t announce any new tools, but it describes the changes the company has made in response to the explosion of AI capabilities. Those include a badge identifying artist identities suspected of being AI-generated, “Verified by Spotify” badges for human artists, and a way to acknowledge the use of AI and specify how it was used in song credits.
Spotify said it noticed listeners were consistently skipping or flagging songs identified as “slop or spam.” It said user complaints decreased after the company made changes to its recommendation system.
Earlier this year, Apple Music, another streaming service, told Billboard that more than a third of monthly uploads to its platform were 100% AI-generated. French company Deezer similarly reported that 44% of its daily uploads were AI-generated tracks.
It seems people don’t love that. In addition to the philosophical issue of whether AI “music” is even music — after all, AI is missing the human element that makes artistry so impressive and allows for human connection across hundreds of miles and Bluetooth connections — there are also obvious concerns about people using AI to copy artists’ voices and infringe on their intellectual property. After all, generative AI has to be trained on something, and that something is mostly human-created work, often without permission. Some artists have even had AI-generated music uploaded under their names, and it appeared on their official pages.
Taylor Swift’s company notably filed several patents earlier in the year, including for her voice saying the words “Hey, it’s Taylor,” perhaps to protect herself from being mimicked with generative AI.
Audrey Schomer, a media analyst and research editor at Luminate who authored a report on AI and music earlier this year, said a study found people had a negative view of AI in music — an attitude that only grew more negative as 2025 progressed. However, about a third of those surveyed felt indifferent towards AI music.
One issue with the flood of AI “slop” — a term emphasizing low-effort, low-quality and high-volume releases that clutter the digital sphere — is its effect on real artists. As NPR reported in May, Spotify, Apple Music and several other companies use a pro rata model, meaning that the percentage of total royalty payouts an artist receives is determined by what percentage of streams on the platform their catalogue comprises. As a result, large volumes of AI music could dilute a human artist’s presence on the platform and lead to lower royalties. An open letter from artists’ groups to AI music company Suno expressed that concern in February.
Spotify said in the press release that AI-generated music is “10 times more likely” to fall below its 1,000-stream minimum for a song to be monetized.
The company said it tries to identify AI-generated content using listener engagement and feedback, upload patterns, metadata, and information from third-party AI detection tools. But it acknowledged that those detection tools aren’t great — and that more sophisticated tech will only make it harder to identify AI content.
Other music streaming platforms have varying policies on AI content.
Apple Music announced a plan to tag AI-generated content with “Made With AI,” with reports saying the feature is set to roll out by the end of the year. Variety reported that record labels and distributors will be required to report songs where “a material portion of the content” is generated using AI.
YouTube Music announced last week that it was rolling out its own labeling system for AI-generated music, saying it was “teaming up with music partners and using our own detection tools to identify AI-generated tracks, applying the same labeling principles already used across YouTube.”
Bandcamp, a platform that attracts many emerging artists for its openness and simplicity and allows users to buy albums at prices set by artists, banned all AI-created music at the start of this year.
“We reserve the right to remove any music on suspicion of being AI-generated,” the company said.
